Now -- how to read this plot:
Along the x-axis is the spectrum number. In this dataset, there are 1168 spectra each taken 2 seconds apart. So this entire dataset took 2336 seconds, or 38 minutes, 56 seconds. Along the y-axis is the pixel location of the detected peaks. So if you can imagine a player piano strip, the song will be played from left to right, with each dot representing a spacial location on the Sun. In this case, the pixel range (0-548) corresponds to a nice audio frequency range, so I didn't transpose or spread out the frequency range in any way. So a pixel location of 432 will correspond to a frequency of 432 Hz. The amplitude of the sound corresponds to the brightness of the pixel value at that peak.
The first plot below shows the peaks above a data value of 25 ADU and as expected quite a few notes are being played at the same time. The 2nd plot shows the peaks above a data value of 50 ADU where there are fewer notes being played. Notice that starting at about spectrum 800 (along the x-axis) and ending about 900, the strip being played starts playing a lot more frequencies. It also turns out that the amplitude of these tones increases by as much as a factor of 300! What is this? Well, it turns out to be IRIS moving through the South Atlantic Anomoly (SAA). In the sound sample, the amplitude is obviously saturated, so I've decreased the volume of that section by 10 dB so your ears aren't blown out.
|Peaks Above 25 ADU|
|Peaks Above 50 ADU|
And here are a couple closeups. The first is from the beginning of the dataset. The 2nd is right before and through most of the detection of the SAA. The closeups make it a bit easier to understand what's being played.
I should mention here that these sound samples SHOULD NOT BE PLAYED THROUGH A SPEAKER SYSTEM. Frequencies and amplitudes might be too much for your system which could result in permanent damage. Listen to this stuff with headphones and make sure you have control of the volume.
Here is the playlist:
IRIS Spectra Player Piano